Microwave photonics instantaneous frequency measurement scheme assisted by machine learning method

This study proposes a microwave photonics instantaneous frequency measurement (IFM) system assisted by machine learning , employing a Polarization-division multiplexing dual-parallel Mach-Zehnder modulator (PDM-DPMZM) as the primary modulation device. The system utilizes the periodic attenuation of dispersion-compensating fibers (DCF) to establish an intensity ratio function (ACF), enabling large-range and high-precision frequency measurement. Additionally, the Stacking ensemble learning method from machine learning is employed to train and test stimulation data obtained from the measurement system, further enhancing the accuracy of frequency measurement. The simulation results indicate that the proposed IFM system achieves an error within 20MHz, removing a small amount of abnormal data, within the frequency range of 4.5GHz to 20GHz. This study provides a novel approach and methodology for the design and optimization of microwave photonics IFM systems, offering practical application value.